By contrast, in interspecific hybrids between M. ... mtDNA is limited to interspecific crosses, which rarely occur in ... Days of pregnancy were counted from the.
Proc. Natl. Acad. Sci. USA Vol. 92, pp. 4542-4546, May 1995 Evolution
Elimination of paternal mitochondrial DNA in intraspecific crosses during early mouse embryogenesis HIDEKI KANEDA*, JUN-ICHI HAYASHIt, SUMIYO TAKAHAMA*, CHOJI TAYA*, KIRSTEN FISCHER LINDAHLS, HIROMICHI YONEKAWA*§
AND
*Department of Laboratory Animal Science, The Tokyo Metropolitan Institute of Medical Science, Honkomagome, Bunkyo-ku, Tokyo, 113 Japan; tInstitute of Biological Sciences, University of Tsukuba, Tsukuba, Ibaraki, 305 Japan; and tHoward Hughes Medical Institute, Departments of Microbiology and Biochemistry, University of Texas Southwestern Medical Center, Dallas, TX 75235-9050 Communicated by John C. Avise,
University of Georgia, Athens, GA, January 6,
1995
(received for review July 28, 1994)
cifically amplify paternal mtDNA from a single mouse embryo in early development, such as the pronucleus or two-cell stage.
To examine whether mtDNA is uni- or bipaABSTRACT rentally transmitted in mice, we developed an assay that can detect sperm mtDNA in a single mouse embryo. In intraspecific hybrids of Mus musculus, paternal mtDNA was detected only through the early pronucleus stage, and its disappearance coincided with loss of membrane potential in sperm-derived mitochondria. By contrast, in interspecific hybrids between M. musculus and Mus spretus, paternal mtDNA was detected throughout development from pronucleus stage to neonates. We propose that oocyte cytoplasm has a species-specific mechanism that recognizes and eliminates sperm mitochondria and mtDNA. This mechanism must recognize nuclearly encoded proteins in the sperm midpiece, and not the mtDNA or the proteins it encodes, because sperm mitochondria from the congenic strain B6.mtsPr, which carries M. spretus mtDNA on background of M. musculus (B6) nuclear genes, were eliminated early by B6 oocytes as in intraspecific crosses. We conclude that cytoplasmic genomes are transmitted uniparentally in intraspecific crosses in mammals as in Chlamydomonas and that leakage of parental mtDNA is limited to interspecific crosses, which rarely occur in
MATERIALS AND METHODS Mouse Strains. C57BL/6J (B6) were purchased from CLEA Japan (Osaka) and inbred M. spretus was from The Jackson Laboratory. The mtDNA congenic strain DBA/2J-mtJ/Stm (D2.mtJpn) was established by backcrossing females carrying mtDNA of Japanese Mus musculus molossinus (16) from donor strain ddY to DBA/2J males (Mus musculus domesticus). The mtDNA congenic strain B6.mtsPr/Kfl was established in the same way by Wharton from a female M. spretus mitochondrial donor and backcrossed by Wharton (F5), then by Fischer Lindahl (+Fs) (17), and finally by one of us (H.Y.) (+F6), to B6 males (M. m. domesticus). It is now maintained by brothersister mating. The strains are summarized in Table 1. Preparation of DNA from A Single Sperm, Egg, or Embryo Suspension. Cauda epididymis was removed from mature mice and placed in TYH medium (18) to prepare a sperm suspension. Unfertilized eggs were collected from oviducts of mice superovulated by pregnant mare's serum gonadotropin and human chorionic gonadotropin (19). Early stage embryos were collected from the oviduct or uterus of mice that had been mated naturally. Days of pregnancy were counted from the appearance of plugs (day 0) and developmental stages were confirmed by phase-contrast microscopy. A single sperm or single egg was isolated from diluted sperm or egg suspensions with a capillary for microinjection under a phase-contrast microscope, placed in 10 ,tl of distilled water, and stored at -80°C. Preparation of DNA from Neonates or Adult Tissues. Neonates (3 days after birth) were killed by ethyl ether anesthesia and minced in RSB medium (50 mM-HCl, pH 8.0/10 mM EDTA/10 mM NaCl) with a disposable razor blade in a disposable plastic dish. DNAs were isolated by the proteinase K/SDS/phenol method (20), dissolved in distilled water, and used for PCR. Tissues removed from an adult mouse, killed by ethyl ether anesthesia, were minced in RSB medium and DNA was prepared as described above. In Vitro Fertilization with Rhodamine-Stained Sperm. The sperm suspension was incubated at 37°C for 2 hr in 5% CO2 in air to induce capacitation. Rhodamine 123 (final concentration, 0.1 jLg/ml) was added to the activated sperm, which were placed at 37°C in a CO2 incubator for 5 min, washed thoroughly on a Millipore filter, and suspended in TYH medium. In vitro fertil-
nature.
Uniparental, female or maternal, transmission of cytoplasmic genomes has been universally observed throughout all phyla of
animals (refs. 1 and 2 and references therein) and plants (refs. 3-6 and references therein). Active degradation of the chloroplast and mitochondrial genomes from one parent has been proven in lower plants with isogametes, such as Chlamydomonas, and a gene involved in this process has recently been cloned from Chlamydomonas (7). One advantage of uniparental transmission that has been proposed is that it evolved to prevent the spread of deleterious cytoplasmic genomes (8). This concept of uniparental transmission has been challenged by recent findings of biparental transmission of mtDNA in several animal species (9-14).¶ The new evidence suggested that biparental transmission does occur in species in which intact sperm enter the egg cytoplasm during fertilization. In particular, Gyllensten et aL (12) detected paternal mtDNA by the PCR in interspecific mitochondrial congenic mice derived from backcrosses between Mus musculus and Mus spretus. Because the paternal contribution was only 0.01-0.1%, these authors suggested that earlier failures to detect paternal mtDNA were due to low sensitivity of applied assays (1, 15). The situation remains ambiguous, however, because many reported cases of paternal transmission involve inter- rather than intraspecific hybrids (9-14). Considering that matings in nature by definition occur mostly within the species, it is important to examine whether mtDNA is biparentally transmitted also in intraspecific hybrids. To address this question, we have developed a sensitive PCR technique that can spe-
§To whom
correspondence avoid confusion, we define "biparental transmission" used here as ¶To that mtDNA derived from both female and male parents that is should be addressed.
transmitted to both female and male progeny. Thus, this does not phenomenon observed in male progeny of Mytilus that is transmitted by both maternal and paternal lineage of mtDNA (14). Even in that case, however, paternal leakage in female progeny has been observed in interspecific hybrids but not in intraspecific hybrids. mean the
publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in
The
accordance with 18 U.S.C. §1734 solely to indicate this fact.
4542
Evlolution: Kaneda et at: Table 1. Genomic
Proc. Natl. Acad Sci. USA 92
4543
composition of mouse strains used D2.mtJPn
B6
C57BL/6
Nuclear genome
Mitochondrial
(1995)
DBA/2J
M. m. domesticus
M. m. domesticus ddY M. m. molossinus
C57BL/6
genome
M. m. domesticus
Cross with B6
Strain M. spretus/J M. spretus
Intraspecific
Intraspecific
ization was done by adding the rhodamine-stained sperm suspension to eggs suspended in TYH medium. The presence of male mitochondria and the developmental stages were confirmed by fluorescence and phase-contrast microscopy, respectively. Synthesis of PCR Primers Specific for mtDNA Haplotypes. We synthesized two PCR primers specific for each of the four mtDNA haplotypes and four primers common to all four, two of those primers (COM1 and COM2) in the D-loop region (Fig. 1). The other two common primers (OL1 and OL2R, not shown in Fig. 1) span the replication origin of the L strand, which is highly conserved among the four mtDNA haplotypes (unpublished). OL1 corresponds to the sequence from 4985 to 5001 and OL2R was complementary to the sequence from 5275 to 5293 (21). Primers were purified by reverse-phase column chromatography as recommended by the manufacturer (Applied Biosystems). Preparation of Specimen, PCR, and Detection of Product. To avoid contamination by exogenous mtDNA, we developed a technique that used only one reaction tube from mtDNA preparation to PCR amplification. The frozen materials were thawed, and proteihase K was added to a final concentration of 50 ,g/ml and distilled water was added to a final volume of 20 ,ul. The mixture was incubated at 37°C for 1 hr to digest proteins and then at 95°C for 10 min to inactivate proteinase K. Mixtures for PCR were set up with a GeneAmp kit as recommended by the manufacturer (Perkin-Elmer/Cetus) in a final volume of 100 ,ul. The first round of PCR employed the outer, haplotype-specific (DOM1, MTJ1, SPR1, or SPE1) and common (COM1) primers for 30 cycles of 94°C, 1 min for denaturation, 45°C, 1 min for annealing, and 72°C, 2 min for extension. Two sets of mixtures were set up for the second round of PCR, which was run under the same conditions as the first. Each used 1 gJl of the first PCR mixture as template. One set
M.
spretus
Interspecific
B6.mtsP'
C57BL/6 M. m. domesticus M. spretus
Intraspecific for nucleus; interspecific for mtDNA
had inner, haplotype-specific (DOM2, MTJ2, SPR2, or SPE2) and common (COM2) primers; the other used common primers OL1 and OL2R for a positive control. After the second round of PCR, 20 ,pl of the reaction mixture was applied to a gel of 3% NuSieve agarose/1% agarose and electrophoresed in lx TAE buffer. Gels were stained with ethidium bromide (0.1 ,ug/ml) to detect PCR
products. Djrect Sequencing of the Second Round PCR Products. The products amplified by second round PCR were eluted from the gel using an Ultrafree C3HV spin column (Millipore, Japan). Direct sequencing with 32P-end-labeled primers was performed using a dsCycle sequencing system as recommended by the manufacturer (Bio-Rad). RESULTS
Sensitivity of the PCR Method. Preliminary experiments showed that primers had to include at least three substitutions at or near their 3' ends to prevent amplification of maternal mtDNA, which is in at least 1000-fold excess. Sequence comparison in the entire D-loop regions of mtDNA showed that no common inbred strain fulfills this conditions (22, 23). We therefore used mtDNA congenic strains that possess mtDNA from a different subspecies, M. m. molossinus (D2.mtJPn), or a different species, M. spretus (B6.mtsPr), with large nucleotide diversity from M. m. domesticus, which provide the nuclear genome (see Table 1). One round of PCR was insufficient to detect the paternal mtDNA in a fertilized egg. We therefore included a second round with nested primers (Fig. 1). Although each primer set had single-sided specificity, we detected no other products with
* * * * * * 15398 * * * 15497 CTMTTA CTACTTCTGAGTACATATT TACATAGTACMCAGTACATTTATGTATATCGTACATTAMCTATTTTCCCAAGCATATAAGCAGTA A.... ...................C........................
domes t i cus molossinus
spretus/J C57BL/6J-mt r .....................................A....
...................................................................................................
*
15498
DOM 1.
*
*
*
*
*
*
*
*
15593
DOM 2* T C ATC ATA ICATTAAATCAATGGTTIAGGTCATAAAATCAATCCT AAAAATAAAA ACTAATG- TTATAA TAAATTTATMTII 1.*IT MT 2--" A .c. I... T.. ......... ..A. -.. -T .. C...T.... . -C.T... ............. L ............ SPR 2SPR 1 . I T CA . ...... -............ A. -. T........TC ....T..T... TA....... C.
domes t i cus
molossinus
spretus/J
...
.
SPE 1.* .
C57BL/6J-mtSPr
P
CA.....
A -
2-*
-
.C.CC .
...
.A. .-..T.....C. .TC.....T..T..
* * * * * * *15693 --COM 2 GGACATATCTGTGTATCTGACATACACCATACAGTCATAAACTC CTCTCCATATGACTATCCCC CCCCATTTGGTCTATTAArCTACCATCCTC A.......................................
15594
domfes t i cus
molossinus
spretus/J
C57BL/6J -t
.. ...
15694
*
*
*
*
*
*
*
............
15793
--COi-1 CGTACCAACAACCCGCCCACCAATGCCCTCTTCTCGCTCCGGGCCATTAAACTTGGGGGTAGCTAAACTGAAACT ATCAGACATCTGGTCTT
domes t icus molossinus
spretus/J C57BL/6J-mt s
................................................................................
....................................................................................................
r
....................................................................................................
...................................................................................................
FIG. 1. D-loop region sequences of the four mtDNA haplotypes used in this study [nucleotides 15491-15771 of Bibb et at (21)]. Dots indicate identity with the domesticus sequence and dashes indicate gaps. Primers are marked by long boxes. The recognition sites of restriction endonuclease Mse I (TTAA) within the PCR-amplified region are underlined. Sequences 3' of primers MTJ2 and SPR2 (doubly underlined) were confirmed by direct sequencing of PCR products.
4544
Evolution: Kaneda et aL
Proc. Natl. Acad. Sci. USA 92
Table 2. Detection of paternal mtDNA at several
early embryogenesis
stages of
1
2
(1995) 3
Positive/total no. of embryos B6 mated To
Stage Pronucleus early Pronucleus late Two cell Four cell Morula
Blastocyst
Neonate -, Not tested.
To
D2.mtJpn
M.
spretus 5/7
18/20 0/23 0/9 8/11 0/6 0/3 0/7 -25/45
To B6.mt spr
5/9 0/19 0/6
different molecular sizes in agarose gels and no unexpected bands after Mse I digestion or in sequencing gels (data not shown). Our primers thus have enough specificity to detect paternal mtDNA in very low amounts. Early Elimination of Sperm mtDNA in Intraspecific Hybrid Embryos. To detect paternal mtDNA in early stage embryos from a (B6 x D2.mtJPn) F1 cross by PCR, we used primers specific for the sperm mtDNA (MTJ1 plus COM1 in the first round and MTJ2 plus COM2 in the second; Fig. 1). PCR products of 168 bp were detected in the embryos through the early pronucleus stage as well as in the control that contained an unfertilized egg and a sperm, both of which had been inactivated (Fig. 2). No such PCR product was detected at any later stage, including the late pronucleus, or in an unfertilized egg (Fig. 2, Table 2). To confirm that the 168-bp product was derived from sperm mtDNA, diagnostic polymorphisms were checked by direct sequencing and by digestion with Mse I. The PCR product possessed the TGATATAAACCA sequence specific for molossinus mtDNA (Fig. 1, doubly underlined site), showing that it was derived from the sperm. Mse I digestion of the PCR product of molossinus DNA should yield fragments of 88, 53, and 8 bp (Fig. 1), of which we observed the two larger ones, whereas the domesticus products should give 101, 41, 17 and 9
bp (data not shown). When control primers, common to both mtDNA haplotypes, were used, a 325-bp PCR product was detected at all stages of embryo development (Fig. 2). Absence of the 168-bp product was therefore not due to technical failure of the PCR. Thus, it is clear that paternal mtDNA did exist in the egg cytoplasm but disappeared after the early pronucleus stage. To further prove that we were indeed detecting mtDNA from sperm midpiece that had entered the cytoplasm rather than remained stuck to the
zona
pellucida, we removed the
A
A
B
FIG. 3: Phase-contrast (A) and fluorescence (B) micrographs of a sperm bound on the zona pellucida of an unfertilized egg (1) an embryo at 3-4 hr after in vitro fertilization (2), and a pronucleus stage embryo (3). Sperm, indicated by arrowheads, were stained with rhodamine 123 before fertilization.
forceps (chemical removal might cause degradation of the sperm midpiece). The zone and the embryo were then tested in separate PCR, and paternal mtDNA was detected only in the embryo and not on the zona (Fig. 2B). Inactivation of Sperm Mitochondria. To examine how long sperm mitochondria remain active in the egg cytoplasm, we fertilized eggs in vitro with rhodamine-stained sperm. Mitochondria stained by rhodamine fluoresce as long as they remain intact (24). We could detect a sperm midpiece with rhodamine fluorescence in egg cytoplasm at the stage of second polar body formation (Fig. 3). Phase-contrast microscopy confirmed that whole sperm was in the egg cytoplasm at that stage. The sperm mitochondria thus function normally in the egg cytoplasm before the pronucleus stage. However, the fluorescent signal disappeared at the early pronucleus stage (Fig. 3), suggesting that the sperm mitochondria became inactivated, in accordance with our PCR results. Persistence of Paternal mtDNA in Interspecific Hybrid Embryos. To confirm biparental transmission in interspecific hybrids (12), we examined the behavior of paternal mtDNA during development of (B6 x spretus)F1 embryos. Single embryos of each developmental stage were subjected to PCR with primers specific for the paternal mtDNA (SPR1 plus COM1 and SPR2 plus COM2), and we detected PCR products of 168 bp in about 70% of early stage embryos (Fig. 4, Table 2). Direct sequencing confirmed that the 168-bp product possessed the AACATACACCC sequence of the M. spretus zona with
2 1 3 4 5 Sc SCSC Sc SCM
B
f*335p
-291 P8 mtD patemal >uc
-210 -162
patemal prorI'~
E ZM
..210bp
"-162
FIG. 2. Detection of paternal, molossinus, mtDNA in intraspecific crosses between B6 and D2.mtJPn (A) Whole embryo: 1, unfertilized egg; 2, unfertilized egg and an inactivated sperm combined; 3, early pronucleus stage embryo; 4, late pronucleus stage embryo; and 5, two-cell stage embryo. Primer pairs for the first round of PCR were MTJ1 and COM1, specific for paternal mtDNA. For the second round, they were either specific (S MTJ2 and COM2) or common (C = OL1 and OL2R). (B) Early pronucleus stage embryo: E, embryo without zona pellucida; Z, zona pellucida; M, size marker (+X174 HincII digest). =
Evolution: Kaneda et at. 1
Proc. Natl. Acad. Sci. USA 92
.335bp
-291 -210 -162
product paternal mtDNAproduct
FIG. 4. Detection of paternal, spretus, mtDNA in the interspecific 1, unfertilized egg; 2, unfertilized egg combined; 3, two-cell stage embryo; and 4, neonate 3 days after birth. S, C, and M are as in Fig. 2, except that primers specific for M. spretus (SPR1 in the first round and SPR2 in the second) were used.
cross between B6 and M. spretus: and an inactivated sperm
site as the sperm mtDNA (Fig. 1, doubly underlined site). The paternal mtDNA was also retained in more than half of F1 neonates (Fig. 4, Table 2). We thus confirmed that paternal transmission does occur in interspecific hybrids but is variable. The Exclusion Mechanism Recognizes Nuclearly Encoded Factors. To examine whether mtDNA is directly involved in the intraspecific recognition of paternal mitochondria, we used a mtDNA congenic strain, B6.mtspr (Table 1). This strain was derived from the same progenitor stock as that used by Gyllensten et at (1, 12, 17). Because all mitochondrially encoded proteins are localized exclusively on the inner membrane (2), the mitochondria of this strain have an outer membrane from M. musculus only. As was the case for intraspecific hybrids, the spretus mtDNA from these congenic male mice could not be detected at any stage from the late pronucleus to neonates, whereas the mtDNA was detectable in the early pronucleus stage embryos (Fig. 5). The paternal factor determining the mode of mtDNA transmission is thus
encoded in the nuclear and not in the mitochondrial genome.
DISCUSSION Exclusion of Paternal mtDNA in Intraspecific Crosses. We have demonstrated that paternal mtDNA is eliminated from the cytoplasm of intraspecific hybrid embryos of the mouse at the late pronucleus stage. The disappearance of sperm mtDNA is not due to lack of sensitivity of our technique, which 1 SC
2
3
4
5
SCSCSCSCM
common mtDNA-
product paternal mtDNAproduct
e335Wp -291 -210 -162
FIG. 5. Detection of paternal, spretus, mtDNA in the intraspecific 1, unfertilized egg; 2, unfertilized egg before fertilization; 3, early pronucleus stage embryo; 4, late pronucleus stage embryo; and 5, two-cell stage embryo. S, C, and M are as in Fig. 2, except that primers specific for M. spretus (SPE1 in the first round and SPE2 in the second) were used. crosses between B6 and B6.mtsPr: and a sperm that was inactivated
4545
paternal mtDNA in a single embryo at the early pronucleus stage (Figs. 2, 4, and 5). Paternal mtDNA would not be diluted by selective replication in the early embryo, because the mtDNA content remains constant from the onecell to the blastocyst stage (25). Furthermore, we did detect paternal mtDNA in 54-72% of interspecific F1 hybrid embryos, even in neonates (Fig. 4, Table 1). If mtDNA was indeed biparentally transmitted within the species as proposes (12), then one should expect heteroplasmic individuals with multiple base substitutions or multiple haplotypes to be more frequent. Several attempts to find such individuals in backcross experiments (1, 12, 14, 26) or by examination of naturally occurring hybridogenic populations (9, 11, 13, 27-29) have succeeded in detecting paternal mtDNA only in interspecific hybrids (9-14) but never in intraspecific hybrids (1, 26-29). Of course, this does not exclude the possibility that biparental transmission can occur in intraspecific hybrids. While this may be true in the mouse studies (1, 12), biparental transmission was detected in interspecific, but not in intraspecific, hybrids of Drosophila (10) and Mytilus (11, 13, 14, 29) by Southern blot hybridization or even ethidium bromide staining (9). Our finding in the mouse that paternal mtDNA is eliminated during early embryogenesis (Fig. 2) may explain the general failure to detect the leakage in intraspecific can detect
2 3 4 SC M C S SC
common mtDNA-
(1995)
cross. Active
Degradation of Sperm Mitochondria. We suggest that maternal transmission of mtDNA is due to active degradation of paternal mitochondria, signaled by loss of their membrane potential in the egg cytoplasm of early stage embryos (Fig. 3). The following lines of evidence support this conclusion. Hiraoka and Hirano (30) examined the fate of the sperm midpiece in the cytoplasm of hamster eggs by electron microscopy. During the two-cell stage, numerous multivesicular bodies gather around the midpiece and fuse with the sperm mitochondria, whereupon the mitochondria are degraded and digested by the multivesicular bodies. Heteroplasmic mice with stable maternal transmission can be created by intraspecific embryo fusion (31), but when mitochondria from liver or testis are microinjected into mouse zygotes, the exogenous mtDNA cannot be detected in the progeny (32), suggesting that non-oocyte, exogenous mitochondria are excluded from egg cytoplasm. Species Specificity of the Exclusion Mechanism. In interspecific hybrids, we observed variable paternal transmission of mtDNA (Table 2), confirming previous reports (9-14). In particular, our findings are consistent with the results of Gyllensten et at (12), but our experimental design produced a different conclusion. Hoping to accumulate paternal mtDNA through successive backcrossing, these authors examined tissues from adult mice after 8-26 generations of backcrossing between M. musculus and M. spretus. They consistently found 0.01% and sometimes 0.1% of paternal mtDNA with uniform tissue distribution patterns in their congenic mice. On the other hand, our B6.mt spr strain, derived from the same original stock, showed no such heteroplasmy when its unfertilized eggs were subjected to PCR (data not shown). In experiments that involve repeated backcrossing, the females become progressively more paternal type with respect to their nuclear genome, whereas their mitochondrial genome remains maternal type. With each succeeding generation, it thus becomes less likely that introgressing paternal mtDNA will escape the exclusion mechanism and increase heteroplasmy. In other words, it is unlikely that repeated backcrossing actually enriched paternal mtDNA. It follows that the congenic mice used by Gyllensten et at (12) must have retained a heteroplasmy that originated in the earliest generation(s). Once established, heteroplasmy could be maintained by uniparental (maternal) transmission through successive generations (12).
4546
The tissue distribution of paternal mtDNA in (domesticus x spretus)F1 hybrids can address this question. In our preliminary examination, paternal mtDNA was randomly distributed; paternal mtDNA was found in a given organ in 10-30% of the F1 mice and in at least one organ on >50% of the F1 mice (H. Shitara, unpublished), consistent with the results with whole embryos (Table 2). No sex difference was observed on the presence of paternal mtDNA in the F1 hybrid mice, suggesting that the paternal mtDNA is transmitted to female and male progeny with equal frequency. We found 3 of 11 F1 females that possessed paternal mtDNA in their ovaries, suggesting that the frequency of heteroplasmic eggs would be
